g., including FRM from warmer rather than cooler environments) and (iii) a low proportion of FRM from distant populations that are ecologically diverse. Breed et al. (2013) go one step further by presenting a provenance selection decision tree whereby decisions are based on evidence and confidence limits surrounding climate distribution modelling, and the degree of population genetic and/or environmental difference between populations. They propose an admixture provenancing approach for situations where there AZD2281 is high confidence in substantial climatic change, but where little is known about the G × E interaction. In admixture provenancing, seed collection is focused on capturing a wide selection

of genotypes from large populations occurring in various environments, with no spatial bias towards the revegetation site and no regard to gene flow dynamics (Breed et al., 2013). In some cases, habitat conditions will be altered to such an extent by climate change and interacting factors such as land use changes that deliberate movement of FRM along environmental gradients may be necessary (Aitken et al., 2008 and Sgrò et al., 2011). In different countries around the world, provenance transfer or assisted migration approaches are already being integrated

in restoration and tree planting practices. In Western Canada, for example, a forest regulation has been changed to accommodate new seed transfer rules to better match seedlings to expected Selleckchem GSKJ4 future conditions. Ideally FRM transfer decisions should be based on solid field trial data, and provenance transfer planning based purely on climate distribution modelling approaches Angiogenesis inhibitor is still highly controversial (e.g., Seddon, 2010 and Sgrò et al., 2011). This is due to the uncertainties associated with both species distribution models and future climate models (for a discussion see Alfaro et al., 2014, this special issue). In situations where no provenance trial data are available, the composite or admixture provenancing described above may be the more prudent approach

(Breed et al., 2013). Species distribution modelling can nonetheless be useful as a first step for guiding the choice of the potential seed sources to be used in the mixtures described above. This is particularly the case when distribution models are used in combination with genetic characterization data, which can provide complementary information about the genetic diversity profiles among and within source populations (Soldati et al., 2013 and Azpilicueta et al., 2013). Recent advances in geospatial modelling and the proliferation of ever-cheaper genotyping techniques make it possible to better design restoration efforts at the landscape level, not only for matching FRM to present and/or future site conditions, but also for optimising connectivity of populations (McRae and Beier, 2007).